Ceramic-metal joints

JOINING CERAMIC COMPONENTS CERAMIC-CERAMIC AND CERAMIC-METAL JOINTS... 

Rapid fluid flow cannot be achieved with active metal brazes because of the need to form solid wettable reaction product layers for their liquid fronts to advance. Equations (10.1) to (10.2) relating liquid flow rates to the opposed effects of surface energy imbalances and of viscous drag are not relevant. Actual penetration rates are so slow, usually of the order of 1 pm.s, that the usual practice is to place the active metal braze alloy within the joints rather than expecting it to fill them, and, as explained already, gap width is not the dominant consideration when designing ceramic-metal joints. 

Such excellent or at least adequate capillary behaviour is also typical of the process variant known as eutectic bonding in which the transient creation of a liquid phase is caused by the interdiffusion of two chemically different metal alloy component materials. In the laboratory variant process known as partial transient liquid phase bonding, (Shalz et al. 1992), a coated interlayer is used for ceramic-ceramic or ceramic-metal joints. In this process the interlayer is a ductile metal or alloy whose surface is coated with a thin layer of a lower melting temperature metal or alloy, for example Ni-20Cr coated with 2 microns of Au. The bonding temperature is chosen so that only the coating melts and the ductility of the interlayer helps to accommodate mismatches in the coefficient of thermal expansion of the component materials. 

The DB-procedure was optimised in respect with the kinetic requirements and the high-temperature mechanical properties of the Ni-superalloy. From the kinetic point of view, the bonding temperature should be over 1000°C when alumina and transition metals are directly bonded [6]. The bonding procedure was always carried out in high vacuum, better than 2-10 mbar (0.2 mPa). The typical thermal and axial compression cycles are presented in Fig.la. It was experimentally found that the ambient bonding temperature is 1100"C or less due to the fast creep of the superalloy beyond this. The compression for the tests was selected as 10 MPa in ceramic-metal joints and 20 MPa in ceramic-ceramic joints [6]. 

D. Munz, M. A. Sckuhr and Y. Yang, "Thermal Stresses in Ceramic-Metal Joints with an Interlayer," J. Amer. Ceram. Soc., 78 [2] 285-290 (1995). 

R. L. Williamson, B. H. Rabin and G. E. Byerly, "FEM Study of the Effects of Interlayers and Creep in Redueing Residual Stresses and Strains in Ceramic-Metal Joints," Composites Eng., 5 [7] 851-863 (1995). 

A model due to Eager et estimates the strain energy in the ceramic in well-bonded ceramic-metal joints. For a small CTE mismatch between the ceramic (C) and the metal substrate (M), but with a large CTE mismatch between the interlayer (I) and the base materials, the elastic strain energy, Uec. >i> the ceramic for a disc-shaped joint is calculated in terms of the yield strength (oyi) of the braze, the radial distance from the center of the joint, and the elastic moduli of the ceramic (Ec) and braze (Ei). Eager et al proposed analytical expressions to calculate Uec as asymptotic approximations (to 1% accuracy) to their finite element calculations these analytical expressions (equations [l]-[3] in ref ) are used here to estimate the strain energy. 

The ceramic membrane typically must also be joined to a metal manifold to collect the product gas. The requirements of ceramic-metal joints are extremely stringent due to the large pressure differential across the membrane. Even very small leaks will create a drop in product purity. The joints are usually at the operating temperature, although cold seals are occasionally considered. The thermal expansion coefficient of the ceramic needs to match that of the metal manifold. 

Two strain and temperature dependent material rTX)dels (in a format acceptable for finite element input) were developed for Palniro 4 (30Au-34Pd-36Ni) and the Au-5Ni-2Pd braze alloys. These models have been added to the material model library developed to explore different systems in the ceramic-metal joint. 

The ability to compare the measured joint strength to the predicted residual stresses will allow the FEA model of the joints to be refined and will provide a tested method to aid in designing ceramic-metal joints in the future. 

The residual stresses in the joint are determined by FEA. The stresses ozz and age are the maximum principal stresses taken from the free surface in the vicinity of the ceramic-metal joint (i.e., ai and 03 respectively). These stresses can be converted into a maximum shearing stress for comparison with the measured joint strength by a Mohr s circle analysis. 

Two Nl-based superalloys (Inconel 600 and 625) were studied for braze alloys In ceramic-metal joints for high temperature applications. These alloys were selected based on their high-temperature performance and ductility. 

Lugscheider, E., Krappitz, H., Peters, R., 1987. Methods for preparing brazed ceramic-metal joints in order to determine the phase composition and joint-structure. Prakt. Metallogr. 24, 195. 

B. Kuhn, F. Wetzel, J. Malzbender, R. Steinbrech, L. Singheiser, Mechanical performance of reactive-air-brazed (RAB) ceramic/metal joints for solid oxide fuel cells at ambient temperature. J. Power Sources 193(1), 199-202 (2009)... 

---